Thallium bromide (TlBr) is a promising semiconductor for use as a direct g-ray detector for nuclear medicine applications, especially for positron emission tomography (PET). It has the potential to provide very high detection efficiency, excellent spatial resolution, low noise and high-energy resolution. TlBr has high-Z constituents, high density, a wide semiconducting bandgap, and based on recent progress with material purification and growth, can be produced with good charge transport parameters for electrons. A compact semiconductor detector operating at room temperature with high spatial resolution, high detection efficiency and good energy resolution would offer unique advantages over LSO, BGO, and CdTe/CZT, which are the most common g-ray detector materials in small animal PET applications. Indeed, TlBr has better stopping power at 511 keV than any other material currently used in PET scanners. High performance PET detector modules could be utilized to build a scanner able to image substructures within the mouse brain, or to quantitatively assess tumor heterogeneity and microenvironment in a mouse cancer model. We believe that this proposal, which will develop the world's first TlBr-based PET detector modules, will be able to bring this promising technology to a point where its commercial exploitation will be possible. The goal of this proposal is to develop two high-sensitivity TlBr detector modules and associated electronics with intrinsic spatial localization of 0.5 mm in two dimensions and 2.5 mm in the third (depth) dimension. The specific aims address steps needed to produce detector modules from existing high quality detector grade TlBr ingots grown by the travelling molten zone method. The ingots are large enough (30 mm and 50 mm diameter) to cut substrates from which monolithic 20 mm x 20 mm detectors can be fabricated. We will: (1) optimize material processing for planar detector fabrication; (2) produce and test single cross-strip detectors; (3) implement multiplexed multi-channel readout and characterize detector performance in that configuration; (4) build two modules of stacked detectors and use them in a benchtop system to evaluate PET imaging performance. Monte Carlo simulations, incorporating experimentally-measured detector data, will also be used to predict the performance of complete small-animal PET scanners based on these TlBr detector modules. These aims will provide the necessary developments in process and design to produce usefully large-area and high-performance detector modules made from TlBr, aimed initially at providing the sub-mm spatial resolution and high detection efficiency needed for small-animal PET applications.